Progressive Consumption Taxes

Ruy T. has an interesting post that got me, once again, thinking about consumption taxes. At first glance, consumption taxes appear intrinsically regressive because poor and middle class households spend most of their income on consumption while wealthy households spend only a fraction of their income on consumption. Suppose there’s a consumption tax of 20% and that poor households spend 90% of their income on consumption (food, clothing, shelter, and miscellany) while wealthy families spend only 50% of their income on consumption. In this example, 18% of the poor household’s income goes to taxes while only 10% a wealthy family’s income goes to taxation.

But a system of exemptions could, in principle, make consumption taxes progressive. Exempting spending on housing from taxation is a good starting point because as income rises, a smaller portion of income is spent on housing. Continuing the example, suppose the poor household spends 40% of its income on rent and the wealthy family spends 20%. The taxable consumption by the poor family then falls to 50% of their income and the wealthy family’s taxable consumption falls to 30% of income. So with a 20% consumption tax, the poor family pays 10% of its income in taxes and the wealthy family pays 6% — still regressive, but less so. To make it actually progressive, high taxes could be added to luxury items, and further exemptions (e.g., of staple foods) could be implemented. Of course, for every exemption implemented, the tax rate on the non-exempted items has to be increased if the exemption is to be revenue neutral.

My fear is that a progressive consumption tax would lead to massively distorted relative prices, reducing economic efficiency (causing a “dead weight loss” in Econ 101 terminology). Worse, the lobbying and influence activity by firms seeking to have their goods and services favorably taxed would be severe. Even worse, development and design decisions would be distorted towards creating products likely to receive favorable tax treatment, creating further inefficiency.(*)

In short, a regressive consumption tax would be a simple, but bad, policy. A progressive tax might be good policy, but the devil is in the details; and in this case, the details are likely to swamp the otherwise potentially positive aspects of a progressive consumption tax.

What about approaching consumption taxes from the other side, by exempting savings? Because all the money you make has to go somewhere, taxing consumption and not taxing savings work out the same. Intuitively, if savings are exempted from taxation then the only money left to tax is income spent on consumption — a consumption tax (a point I made earlier). And this is precisely what the Bush administration’s proposed Lifetime Savings Accounts would accomplish.

This raises three questions: (1) Are the current Bush proposals in fact regressive? (The answer is yes); (2) Is any plan that uses tax incentives to encourage saving necessarily regressive? (The answer is no); and (3) Would a progressive savings-based tax plan also be unwieldy and introduce substantial economic distortions similar to those that would occur under a progressive tax on consumption? (The answer is perhaps not.)

The new Center for American Progress, a center-left think tank, now has a section devoted to this issue. It includes “Critiques of the Bush administration plan” (documenting point (1) above) and “A Progressive Alternative: The Universal 401(k)” (focusing on points (2) and (3) above). Soon, I’ll review the CAP proposal, but in the meantime, you can check it out for yourself.


(*) Don’t income taxes also introduce distortions? Yes, but they only mess up one tradeoff, the labor vs. leisure decision. Moreover, the impact of income taxes on the amount of work people do is ambiguous. On the one hand, it will discourage work because whenever you tax something, people will do less of that thing. On the other hand, it will encourage work because people have to work more to afford a given set of goods. (For another Econ 101 flashback, the first effect is the substitution effect and the second is the income effect). If the two effects are roughly equal in magnitude then the amount people will work when income is not taxed and the amount they work when income is taxed are about the same, meaning that the economic distortion and inefficiency created by an income tax is modest. I doubt that the two effects precisely counteract each other in this fashion, but this does attenuate the distortion; with consumption taxes, there is no similar mitigating effect.